Bingyao Wang

AI / Search / Full-Stack Engineering

Bingyao Wang

Technical leader building AI systems, search products, and agentic tools.

Tech Lead at TikTok Search, former engineer at Meta and Lark. I build AI-powered product systems across visual search, LLM applications, backend infrastructure, and integrations with a strong focus on quality, speed, and practical business impact.

View ProjectsResumeGitHubEmail

Building end-to-end AI products with technical depth, product clarity, and founder execution.

About

A concise view of the technical systems and product surfaces I focus on.

I lead and ship AI products end-to-end: from model strategy and retrieval design to backend implementation and product rollout. My work spans LLM systems, search and recommendation, multimodal understanding, and scalable platform engineering. I operate comfortably from 0-to-1 prototyping through production hardening.

Experience

Product-facing engineering across AI systems, search, and platform architecture.

TikTok

Tech Lead Software Engineer, TikTok Search

Current
  • Led visual search initiatives that connect frame-level video understanding to interactive discovery experiences.
  • Built Creator Search Insights capabilities to translate search demand into actionable creator guidance.
  • Developed semantic video tagging and retrieval pipelines to improve ranking quality and search relevance.
  • Scaled LLM infrastructure patterns for internal product teams across inference, evaluation, and integration.

Lark

Software Engineer / Tech Lead

Formerly
  • Built GTM and SEO-aligned growth systems with clear business and product feedback loops.
  • Led development of third-party integration platform capabilities for enterprise workflows.
  • Implemented SaaS connectors and automation primitives for authentication, synchronization, and event processing.

Meta

Software Engineer, Notifications ML

Formerly
  • Worked on ranking and delivery systems for high-scale notification surfaces.
  • Improved infra efficiency for model-related services while maintaining strict reliability and latency targets.
  • Contributed to large-scale ML production systems tied to measurable product engagement metrics.

Featured Projects

Selected systems with strong product framing and technical depth.

Visual Search Pause Tag

A 0-to-1 interactive visual search feature enabling users to pause a video, tap frame entities, and retrieve related content in real time.

Visual SearchVideo UnderstandingRanking

Intelligent Creator Tips

An LLM-powered creator guidance system combining retrieval, query mining, and generation to deliver topic insights, title suggestions, hashtags, and script drafts.

LLMRAGProduct AI

Data360

A third-party SaaS integration platform supporting authentication, data synchronization, and scalable webhooks for enterprise workflow automation.

IntegrationsBackendWebhooks

Instant SEO Coze Bot

A multi-agent SEO content generator using workflow orchestration, domain knowledge retrieval, and structured prompting to produce optimized blog drafts.

AgentsSEOAutomation

BiteSafe

A practical AI side project focused on high-signal product UX and reliable behavior for real-world user decision support.

iOSAI ProductConsumer

AI Travel Planner

An agentic planning experience that combines contextual retrieval and itinerary generation for actionable trip planning.

AgentsContext EngineeringPlanning

Poker Rival

A poker training and analysis platform supporting both head-to-head and multiplayer games, configurable opponent play styles, and three-layer guidance from legacy strategy, platform recommendations, and pro-player insights.

Game AnalyticsLearning ProductDecision Support

Lingo Adventure

An adaptive reading experience that intelligently adjusts material difficulty to each reader level to improve comprehension and progression.

Adaptive LearningPersonalizationEdTech

Publications

Research work focused on faithful summarization and activity recognition.

CONFIT: Toward Faithful Dialogue Summarization with Linguistically-Informed Contrastive Fine-tuning

NAACL 2022

Approaching the Real-World: Supporting Activity Recognition Training with Virtual IMU Data

IMWUT 2021

Skills

Core stacks and capabilities used to build production-grade AI products.

AI / LLM

OpenAIAzure OpenAIPrompt EngineeringSFTRAGAgentsTool Calling

Languages

GolangPythonJavaC++ScalaSQLSwiftJavaScript/TypeScript

Frameworks / Platforms

ReactNext.jsNode.jsTemporalKubernetes

Data / Infra

MySQLMongoDBHiveVector RetrievalScalable Pipelines

Contact

Open to conversations on AI product engineering, technical leadership, and founder-minded roles.

Emailbingyaowang2023@gmail.comGitHubhttps://github.com/BennyWnjWebsitehttps://bennywnj.github.ioLinkedInhttps://www.linkedin.com/in/bingyaow/